XENO+ ™ Nano ML Module

XENO+ WiFi+BLE Nano ML Module is a solderable module and can be used as core ML (Machine Learning) module of new Battery powered ML edge devices of customers so that customers can focus only on adding audio Mics, I2C based sensors and Battery power circuits around this ML module for building their new ML edge devices in short time. This XENO+ Nano ML module is suitable for Always-On-Voice (AOV) and AI for IOT application requirements.

The module has smallest possible size with secure IoT MCU of Nuvoton ARM Cortex-M23 (M2354) Trust Zone series MCU @ 96MHz, 1MB Flash, 256KB SRAM, UART/SPI/I2C ports and GPIOs. This module has USB Type C based 5V Power input with serial debug port and Battery power input options. This module has 3 pin SWD pins for SW development and SW debug via KEIL IDE for Embedded SW development for the device by the users.

This Nano module is suitable for Trusted Execution Environment (TEE) with Trusted Applications (TAs). The key security features are,

  • Tamper-resistant key storage in Flash and SRAM,
  • TrustZone for Armv8-M Technology,
  • 8 regions MPU_NS (for normal world) and 8 regions MPU_S (for secure world),
  • Hardware Crypto Accelerators (AES, ECC and RSA), CRC calculation unit,
  • Up to 6 tamper detection pins and
  • Arm Platform Security Architecture (PSA Certified Level 2 /Level 3) supported.
This module has Syntiant NDP120 neural decision processor that runs multiple Audio and Sensor ML applications simultaneously with minimal power consumption. The NDP120 is designed to natively run multiple Deep Neural Networks (DNN) on a variety of architectures, such as CNNs, RNNs and fully connected networks up to 256 layers. The NDP120’s Audio-0 (AUD0) port is interfaced with on-board stereo PDM Mics and thus helps Customers to readily run and test the Audio ML classification applications on this module. The Audio-1 (AUD1) and I2C interface pins are available in the 20x20 castellated header pins so that Customers can build carrier board to interface with external Audio Mics and I2C sensors respectively.

This module includes InnoPhase IoT’s Talaria TWO™ certified WiFi and BLE module INP1014, that provides an integrated WiFi b/g/n and BLE 5.0 with a single PCB antenna. The WiFi is a single-band, 2.4 GHz, 802.11 b/g/n solution with WPA2/WPA3 personal and enterprise security certified and capable of delivering 65 Mbps PHY rate. The integrated BLE with Long Range, Extended ADV, and 2 Mbps PHY can be used for device provisioning. The Low Power INP1014 module offers a 50% reduction in cThe platform includes InnoPhase IoT’s Talaria TWO™ certified WiFi and BLE module, INP1014, that provides an integrated WiFi b/g/n and BLE 5.0 with a single PCB antenna. The WiFi is a single-band, 2.4 GHz, 802.11 b/g/n solution with WPA2/WPA3 personal and enterprise security certified and capable of delivering 65 Mbps PHY rate. The integrated BLE with Long Range, Extended ADV, and 2 Mbps PHY can be used for device provisioning. The Low Power INP1014 module offers a 50% reduction in current consumption compared to the competition and is suitable for building high bandwidth, battery-operated solutions.urrent consumption compared to the competition and is suitable for building high bandwidth, battery-operated solutions.
Nano ML Module Overview

The below table shows the brief overview of modules:


XNO-W102N-M1 XNO-W112N-M1
Module Image dimensions dimensions
Wireless Interface WiFi 2.4 b/g/n and BLE5.0 WiFi 2.4 b/g/n and BLE5.0
Antenna PCB Antenna UFL Antenna
Audio On-board Stereo PDM Mics On-board Stereo PDM Mics
Sensor Interface a) NDP120 supports external I2S/PDM/TDM and I2C.
b) Host MCU supports UART, SPI, I2C, ADC, DAC, PWM and GPIOs		 a) NDP120 supports external I2S/PDM/TDM and I2C.
b) Host MCU supports UART, SPI, I2C, ADC, DAC, PWM and GPIOs
Pins 20x20 Castellated Pins 20x20 Castellated Pins
Size in mm 60 x 38 mm 60 x 38 mm

Product Features and Specifications:

The XENO+ ™ Nano ML module product features and specifications are listed below:

Table-1: The Product features and specifications


SI.No Features Description
Electro – Mechanical specification
1 Boards Mounting 20x20 Castellated Pins
2 Wired interface 1x USB Type C for 5V Power and Serial debug UART interface for debug messages and user inputs.
3 Antenna Chip antenna or uFL connector for external antenna supports 2.4G
4 User SW programming 3 pin SWD pins for SW programming and debug via KEIL IDE
5 Operating temperature -40 to +85 ℃
Note: The NDP120 is qualified from 0 to +80 ℃.
6 Operating humidity 95% or less
7 Size 60x38 mm
8 Weight 5 grams
Power Specifications
1 Module input voltage 5V from USB Type C connector
* MOSFET based switch to auto cut-off battery power when USB 5V is present.
2 Battery input voltage 2.5V to 5V Battery power from non-rechargeable (or) rechargeable battery. The battery options are listed below:
* Two 1.5V AA Type Alkaline/Drycell non-rechargeable batteries connected in series for non-restricted transport.
* 3.6V AA type (Li-SOCl2) non-rechargeable battery for Industrial applications.
* 4.2V LiFePO4 18650rechargeable battery. The recharging circuit to be added in carrier board by customer.
WiFi/BLE Specifications
1 WiFi 2.4GHz b/g/n * PCB Antenna or uFL connector for external Antenna
2 BLE BLE5.0 with 2Mbps PHY, LE Coded and Extended Advertising
3 WiFi/BLE Data interface UART based AT command interface
4 WiFi Active Mode Power 31mA (Rx Mode @ 1Mbps 802.11b), 178mA (Tx Mode @ 1Mbps 802.11b +17.5dBm)
5 Wi-Fi Power Save Mode Power at 802.11b, 1Mbps (Clean Environment, @3.3V) 57uA (DTIM10)
6 Deep Sleep Mode Power (@3.3V, Memory Retained) 11-19uA (RTC, memory retained, depends on amount of SRAM retained)
7 Module Standby Power X (STM32L4 in standby and WiFi module is in shutdown mode)
8 WiFi Power Shutdown WiFi Power can be shut down by MCU GPIO control signal.
9 Antenna PCB Antenna (or) UFL Antenna Connector
Neural Decision Processor Specifications
1 Neural Processor Syntiant Core 2 NDP120
2 Supported Layers Fully connected, 2D Convolution, Deep-wise convolution, Recur-rent neural network including LSTM and GRU and Average & max pooling.
3 Neural parameters Upto 896K neural parameters in 8-bit mode, 1.6M parameters in 4-bit mode and 7M+ parameters in 1-bit mode
4 Audio interfaces(onboard) 1x on-board Stereo PDM Mics connected to AUD0 port.
5 Audio interfaces(external) 1x Audio interface (AUD1 and AUD2) for external Mics.
  • PDM mode: AUD1 port alone used with PDM_PCLK0, PDM_PCLK1, PDM_PDAT signals on module header.
  • I2S mode (4-wire): Both AUD1 & AUD2 ports are used with I2S_WS, I2S_CLK, I2S_DIN, I2S_DOUT signals on module header.
    		•
    6 Sensor Interface 1x I2C Interface for connecting any I2C sensors like Accel-erometer, Gyro, Pressure, Temperature,..etc sensors
    7 Other Interface 2x GPIO signals available on module header pins. These 2 GPIOs can be alternatively configured as UART: UART_TX and UART_RX interface.
    CPU & Other Specifications
    1 CPU Nuvoton M2354 series MCU with ARM Cortex-M4 at 96MHz
    2 Flash Memory 1MB with Preassigned FOTA section.
    3 RAM 128KB SRAM.
    4 Additional Flash 8MB Serial Flash on SPI bus.
    5 Sensor Interfaces On Function-A pins:
    1x SPI with two chip selects
    2x UART
    1x I2C
    1x CAN
    2x ADC
    2x DAC
    2x PWM
    5x GPIOs
    6 RTOS FreeRTOS.
    Module Pinouts:

    This module has 20x20 Castellated pins. The Left side 20 pins mapping and Right side 20 pins mapping are listed below. The MCU port pins can be assigned with 2 predefined module functions: Function-A and Function-B. The users can also map custom functions as per MCU native GPIO functions on these pins.


    MCU Pinouts

    Left side 20 pins connector signals:

    SNO MCU Pins Function A Function B
    1 PA10 GPIO_PA10 GPIO_PA10
    2 PA9 TM4_EXT GPIO_PA9
    3 PA8 TM5_EXT GPIO_PA8
    4 PF6 GPIO_PF6 GPIO_PF6
    5 GND GND pin of module
    6 PC7 SPI1_MISO TM0
    7 PC6 SPI1_MOSI TM1
    8 PA7 SPI1_CLK TM4/TM2
    9 PA6 GPIO_PA6 GPIO_PA6
    10 PA5 CAN0_TXD CAN0_TXD
    11 PA4 CAN0_RXD CAN0_RXD
    12 PA1 2C2_SCL GPIO_PA1
    13 PA0 I2C2_SDA GPIO_PA0
    14 PC5 UART4_TXD GPIO_PC5
    15 PC4 UART4_RXD GPIO_PC4
    16 PC3 UART3_TXD GPIO_PC3
    17 PC2 UART3_RXD GPIO_PC2
    18 NRST Active LOW RESET signal to MCU. The Push button also asserts RESET signal to LOW.
    19 GND GND pin of module.
    20 VDD_3V3 Regulated 3V3 supply output from module to other circuits of carrier board.

    Right side 20 pins connector signals:

    SNO MCU/NDP120 Pins Function A Function B
    1 AUD1_PCLK1 NDP_PDM_PCLK1 NDP_I2S_WS
    2 AUD1_PCLK0 NDP_PDM_PCLK0 NDP_I2S_CLK
    3 AUD1_PDAT NDP_PDM_PDAT NDP_I2S_DIN
    4 SCL NDP_SCL
    5 SDA NDP_SDA
    6 AUD2_SDO Unused NDP_I2S_DOUT
    7 GPIO7 NDP_GPIO7 UART_TX
    8 GPIO6 NDP_GPIO6 UART_RX
    9 GND GND pin of module.
    10 PB12 DAC0_OUT SPI0_MOSI
    11 PB13 DAC1_OUT SPI0_MISO
    12 PB14 EADC0_CH14 SPIO_CLK
    13 PB15 EADC0_CH15 SPIO_SS
    14 PA15 GPIO_PA15 GPIO_PA15
    15 PA14 GPIO_PA14 GPIO_PA14
    16 PA13 BPWM1_CH3 GPIO_PA13
    17 PA12 BPWM1_CH2 GPIO_PA12
    18 GND GND pin of module
    19 VBAT Battery supply voltage input to module with 2.5V to 5V range. If Battery is connected, the module works with this battery supply.
    20 VBUS_5V USB 5V supply voltage output from module to other circuits of carrier board. This is USB 5V supply. When USB cable is removed, the module will switch to Bat-tery supply on-the-fly, if battery is connected.


    Module Layout and Dimensions:

    This module layout and dimensions are shown below.

    Module Dimensions (in mm)

    dimensions

    Documentation:


    XENO+ ™ WiFI+BLE Nano ML Module Nuvoton Based Datasheet


    XENO+ ™ WiFI+BLE Nano ML Module Alexa demo video



    Video Description

    This video shows the working of ALEXA wake-word detection demo in XENO+ Nano ML module with on-board Mics. This module uses Syntiant NDP120 Neural Decision Processor for running ML classification model based on on-board Mics data. The WiFi+BLE5.0 connectivity enables the ready data uploading feature to Cloud Server.